AEROACOUSTICS RESEARCH IN EUROPE:

usefultenchMechanics

Feb 22, 2014 (3 years and 8 months ago)

228 views


1

AEROACOUSTICS RESEARCH IN EUROPE:

THE CEAS
-
ASC REPORT ON 2001 HIGHLIGHTS




Authors:

Dr M J Fisher and Dr R H Self

Address:

Institute of Sound and Vibration Research


University of Southampton


Highfield


Southampton


Hampshire SO17 1BJ, UK


Short running
headline:

Aeroacoustics Research in Europe


No. of pages: 40


No. of tables: 0


No. of figures: 24



2

AEROACOUSTICS RESEARCH IN EUROPE:

THE CEAS
-
ASC REPORT ON 2001 HIGHLIGHTS



M. J. FISHER and R. H. SELF

Institute of Sound and Vibration Research, Universi
ty of Southampton,

Southampton SO17 1BJ, UK



This paper summarises some highlights of aeroacoustics research in Europe in 2001,
compiled from information provided to the CEAS Aeroacoustics Specialists Committee
(ASC). The CEAS (Confederation of European A
erospace Societies) comprises the
national Aerospace Societies of France (AAAF), Germany (DGLR), Italy (AIDAA), The
Netherlands (NVvL), Spain (AIAE), Sweden (FTF), Switzerland (SVFW) and the United
Kingdom (RAeS).



1. INTRODUCTION

The role of the CEAS Aer
oacoustics Specialists Committee (ASC) is to serve and support
the scientific and industrial aeroacoustics community in Europe. Here “Aeroacoustics” is
to encompass all aerospace acoustics and related areas.


This paper summarises events in 2001. During th
at year four major programmes of
research, namely RESOUND, RANNTAC, RAIN and DUCAT supported by the
European Union, were concluded and a brief summary of each is therefore included in
Section 2. The extension of that work is funded under three new European

supported
programmes, SILENCE(R), TurboNoise CFD, and JEAN; their objectives are outlined in
Section 3.


The remainder of this article comprises contributions by European researchers as
submitted to the editors. Enquiries regarding all contributions shoul
d be directed to their
principal authors who are identified at the end of each subsection.



3

2. EUROPEAN PROGRAMMES COMPLETED IN 2001

2.1. RESOUND (
R
eduction of
E
ngine
So
urce Noise through
U
nderstanding and
N
ovel
D
esign)

2.1.1.
Introduction


Despite signif
icant progress in reducing aircraft noise over the past thirty years,
further improvements are required if passenger growth is to continue without an increase
in noise exposure around airports Increased bypass ratios have reduced jet noise from the
dominan
t levels of the first jet engines, so that now further progress in reducing aircraft
noise requires research on a broad range of noise sources, including turbomachinery noise
(especially fan noise), combustor noise, jet noise (still very important at aircr
aft
departure) and airframe noise (often the dominant source on aircraft arrival).


RESOUND is a 9 MEuro European research programme launched in 1998 involving
18 partners from industry, research establishments and universities across Europe, with
financi
al support from the European Union. The objective of RESOUND is to acquire the
technology necessary to support the design of derivative and new aero
-
engines with noise
levels that are 4 dB quieter than those of aircraft currently entering service.
Conventi
onally turbomachinery noise reduction is achieved by increasing r
otor/stator
gaps increased to reduce interactions, and by selecting rotor and stator numbers so that
sound from rotor/stator interaction decays rapidly. The RESOUND programme
investigated adv
anced methods of reducing turbomachinery noise, namely

fan noise
reduction through reduced tip speed and rotor/stator design, low
-
pressure (LP) turbine
noise reduction through exit guide vane design, and turbomachinery noise reduction
through active stator

design. Other elements of the programme, including combustor
noise and the control of fan noise by aerodynamic devices, are not presented here.


2.1.2.
Fan noise reduction through reduced tip speed and rotor/stator design


A model fan test programme was c
onducted in the Rolls
-
Royce anechoic chamber at
Ansty in the UK. Far
-
field measurements were taken in the controlled environment of a
large anechoic chamber. The 0.85 metre diameter model fan operating at supersonic fan
tip speeds was driven by a 10 MegaWa
tt electric motor, and installed with a turbulence
control screen to simulate intake flows relevant to the flight case.


Two designs of fan rotor were investigated in RESOUND. Low Noise Rotor 1


LNR1


was acoustically designed with a 15% tip speed reduct
ion using 3
-
dimensional
(3D) Computational Fluid Dynamics (CFD). The rotor was designed to have the same

4

pressure ratio versus mass flow relationship as the datum blade with no reduction in
efficiency and adequate stability margin. The rig was tested in No
vember 1999, and
demonstrated more than 4 dB improvement in rotor tone noise, as well as reductions in
rotor
-
stator interaction tones (corrected to the same gap/chord ratio). Broadband fan noise
benefits were not realised.


LNR2 was designed using 3D CFD t
o match the datum blade pressure ratio versus
mass flow relationship, with the same tip speed as the datum rotor. LNR2 was swept
forward to swallow the shock at the tip and stop it propagating forward as noise. 2
-
4

dB
noise improvement was predicted using
steady and unsteady Noise CFD for rotor
-
based
tone sources. The rig was tested in July 2000 demonstrating more than 4 dB improvement
in rotor tone noise.


In addition to low
-
noise rotor designs, a low
-
noise fan swept outlet guide vane (OGV)
was evaluated.
The OGV axial sweep noise benefits were evaluated, and a 20 degree
sweep selected with 3
-
4 dB benefit on interaction tones predicted. The swept OGV was
aerodynamically designed using 3D CFD Navier
-
Stokes methods to have the same
overall aerodynamic loss as

the datum OGV. The swept OGV was tested in July 2000,
and demonstrated a rear
-
arc tone reduction of more than 4 dB.


2.1.3.
LP turbine noise reduction through exit guide vane design


A number of turbine exit guide vane variants were tested on a LP turbine

rig at MTU,
Germany. The blades were designed to reduce rotor
-
stator interaction noise using CFD
predictions of linear perturbations to non
-
linear steady mean flow.


2.1.4.
Turbomachinery noise reduction through active stator design


Tests were conducted
on two different concepts of active stator design, introducing
either anti
-
noise or aerodynamic disturbances to cancel out the fan noise sources. In one
test, loudspeakers were located on the duct walls between the stators, while in the other
ten of the st
ators in the ring were modified to incorporate piezo
-
actuators


In the tests at SNECMA, both concepts gave significant (~10 dB) overall tone
reduction for low frequencies, with an angular sector of ‘silence’ achieved at high
frequencies.



5

2.1.5.
Conclusi
ons


RESOUND has delivered technology for aero
-
engine noise reduction that is suitable
for application in the short term to fan rotor and outlet guide vane designs, and LP turbine
exit guide vane designs and in the longer term to active stator designs.


[
By A.J. Kempton, Rolls
-
Royce plc, Derby, UK]


2.2. RANNTAC

RANNTAC is a European Research Project devoted to the study of new low noise nacelle
technologies. Started in January 1998 and completed in December 2000, this project
coordinated by AIRBUS
-
FRANCE
has allowed investigation into both passive and active
concepts.


In the family of passive concepts, the following new absorbers have been studied:
Hollow sphere liner of METRAVIB, SAA concept of DORNIER, Variable cavity depth
sequence liner of University
of Salford, 3DOF concept of SNECMA, bulk and spiralling
liners of NLR. Most of these technologies have exhibited promising results and the
benefit of two of them has been confirmed by engine tests performed by Rolls
-
Royce
D
eutschland
.


Also in the category

of passive technologies the negatively scarfed intake has been
investigated with the aim of diverting noise toward sky rather than ground thanks to a
longer lower lip. Aerodynamic study and scale model tests carried out by Rolls
-
Royce
have allowed a much
greater understanding of the acoustic benefit of this concept, and
have highlighted the aerodynamic difficulties that must be addressed.


Between passive and active means, adaptive liners have been studied with the aim at
highlighting their versatility wit
h respect to the incoming acoustic wave. Different
approaches have been investigated by Ecole Centrale De Lyon in collaboration with
Metravib and Eads CRC based on acoustic or mechanical adaptation respectively.


In parallel of the study of liners, active
noise control systems have been investigated
with special attention given to the actuator positioning, algorithm and actuator
technology. MTU and ISVR have suggested theoretical rules defining the best position
for actuators so that an optimum control is a
chieved. EADS CRC and CNRS/LMA
developed two algorithms based on in
-
duct or far field control. Positioning and
algorithms have been tested at SNECMA's facility and effects on modes have been
measured by DLR. Associated to these configuration studies, actua
tor technologies have

6

also been investigated with the aim at achieving high levels with low bulk / low weight
devices. Bertin, CTTM, Dornier, METRAVIB, Ferroperm, LC
-
EPFL and Cambridge
Concept have studied various concepts the most mature of which were tes
ted at CTTM
facility by LEMA
-
EPFL. Even though some of these concepts have showed very high
performances, the studies devoted to ANC configurations have demonstrated the need to
improve again the output level/dimension ratio.



Example of new passive liner concept

(Hollow sphere concept developed by
METRAVIB).


The Negatively Scarfed Intake

(Scale model flare manufactured by RR).


porous liner
Feed-forward
LMS controller
H
piezo
-electric actuator
grid
Ue

Example o
f adaptive liner concept

(Acoustically adaptation developed by ECL
and METRAVIB).



Example of new ac
tuator for ANC

(Electrodynamic direct radiator of CTTM).


Figure 2.1 Showing examples of RANNTAC technology.


[By Hervé Batard, AIRBUS
-
France]



2.3. REDUCTION OF AIRFRAME AND INSTALLATION NOISE (RAIN)


RAIN is one of the aircraft noise reduction research

projects partly funded by EC
under the Brite
-
Euram fourth framework. The airframe noise is an important contributor
to the total noise level perceived on ground for large aircraft in an approach

7

configuration. This source has to be reduced if the annoyanc
e around airports is to be
lower. The installation effects on the engine noise and particularly its interaction with the
airframe is another area which requires better understanding in order to achieve further
noise reduction. The objectives of RAIN are to

identify the noise source mechanisms, to
develop noise reduction solutions, to generate noise database and to improve/validate
prediction models.


An A340 flight test has been performed to determine the important aerodynamic noise
sources. These are from
landing gears and the high lift devices. A noise source
localisation tool was successfully applied during this test campaign. Detail studies of the
noise sources have been conducted in separate wind tunnel tests. For the landing gears,
the most important r
adiation areas have been identified using full scale tests in the DNW
tunnel. The results have been used to improve noise prediction models and to design
noise reduction treatments. A significant noise reduction has been achieved.

For the wing noise source
s, an A320 and A321 1/11 scaled model tests have been
performed in the CEPRA 19 tunnel. The identification of the noise sources allowed the
development of noise reduction treatments. The efficiency of the treatments has been
validated at model and full sca
le in CEPRA 19 and DNW tunnel respectively. Very good
noise reduction has also been achieved.

The engine installation effects have been studied separately for fan, jet and core
noise. Wind tunnel tests at model scale have been performed for fan and jet noi
se in the
CEPRA 19 tunnel, whilst for core noise the DERA (now QinetiQ) tunnel was employed.
Good databases have been generated. The results obtained have clearly demonstrated a
significant increase of the level of these sources due to the installation on
the aircraft
model.


[By L.C. Chow, Airbus UK, Filton, Bristol, England]



2.4. DUCAT

2.4.1.
Introduction


DUCAT is a European Research Project aiming to develop and validate duct acoustics
models, which with confidence after validation, can be used as in
dustrial design tools for
the optimisation of liners and actuators. The project started in January 1998 and was
completed in December 2000. Main objectives of the project are:



development, extension and validation of various duct acoustics models;


8



redu
ction of the computational effort required by the models to obtain results up
to dimensionless frequencies of 40;



constitution of a firm experimental database for present and future applications;



demonstration of the improved design capability on a low

noise nacelle of a
generic turbofan and assessment of the applicability of the models.


Prior to the development of the models, the industrial partners (Rolls
-
Royce, Rolls
-
Royce Deutschland, Turbomeca and EADS Airbus) raised the industrial specifications
for
the duct acoustics models. In the work programme, it was anticipated that not all aspects
could be addressed in a single duct acoustics model. Therefore a small number of
numerical models has been developed, based on various methods as FEM (NUIG), BEM
(DTU), coupled FEM/BEM (UTC), non
-
linear propagation (ISVR), ray
-
acoustics
(ONERA) and a 2
-
port matrix method for a duct lined with a bulk absorber (KTH). These
models are partially complementary and partially overlapping, which offers the
possibility to f
ind the best modelling for each aspect of duct acoustics. Three validation
experiments were carried out to constitute a database for current and future applications
(UTC, NLR and ISVR). After validation and a liner design study, the range of
applications o
f the models and the restrictions for use as industrial design tools for nacelle
acoustic optimisation was established.


2.4.2.
Modelling and validation


NUIG improved the computational efficiency of their FEM
-
model by using a pre
-
conditioned solution matr
ix, which can be re
-
used for subsequent calculations and which
gives a CPU
-
time reduction of more than 50%. An existing infinite wave envelope
element formulation has been modified to calculate fan noise radiation. Good agreement
between model predictions
and benchmarks and experimental results was found.
Furthermore, a liner design method based on a gradient optimisation routine was
implemented. DTU extended the BEM
-
modelling for uniform flow to a model for
potential flow, first proposed by Astley and Bain

in 1986 for flows at low Mach number.
An improved variant of this model seemed to be considerably more accurate.


UTC developed a coupled FEM
-
BEM model to study the problem of acoustic
propagation and radiation in a non
-
uniform mean flow. The inner domain

close to the
duct with non
-
uniform flow is calculated using FEM. The radiated acoustic field in the
external domain is calculated using BEM (in the project considered as a no
-
flow region).
The fluid is supposed to be perfect, compressible, and isentropic.

The linear acoustic

9

equations are obtained by the perturbation of the mass conservation and momentum
equations, leading to the so
-
called Galbrun equation. The code is developed and validated
with various comparisons as analytical benchmarks and experiment
al results. Good
agreement was found (Figure 2.2).


The development of theoretical models of non
-
linear propagation of tones radiated
forward of supersonic fans has been studied for many years at the ISVR, and this work
has been successfully continued in D
UCAT. A frequency
-
based model has been
developed, so
-
called FDNS (Frequency Domain Numerical Solution), which predicts the
non
-
linear propagation in a lined intake duct. The predictions were compared to
FANPAC data (a previous Brite
-
Euram project) and data

from a turbofan engine test at
Hucknall. Generally, a good agreement between measured and predicted results was
found.


As a high frequency approximation, ONERA developed a no
-
flow ray
-
model as an
alternative to purely numerical methods. One of the object
ives is to demonstrate the
capabilities of the ray model to handle both in
-
duct propagation and far
-
field radiation.
The model allows a coherent or incoherent contribution of the rays both at the source
emission and at the observer in the far
-
field. Furthe
rmore, the model can handle spinning
modes and non
-
uniform liners. The model was validated using analytical benchmarks, for
which good agreement was found.


KTH wrote, validated and documented a numerical code calculating the 2
-
port matrix
for a duct lined

with a bulk absorber in the case of negligible mean flow. The problem
was handled by first analysing an infinite duct of arbitrary cross
-
section with extended
reaction liners. The resulting eigenvalue problem was solved using the collocation
technique. Wh
en the eigenmodes and wave numbers of the lined ducts were known a
mode matching technique was used to solve the transmission properties of the finite duct.
The starting point for this work was a code previously developed at KTH for analysing
the infinite
problem for automotive dissipative silencers. After validation comparisons
were made between the transmission loss of an axisymmetric liner and one in which the
depth varied azimuthally. The axisymmetric case, with constant liner depth, was predicted
to gi
ve the higher transmission loss.


2.4.3.
Experiments


NLR in September 1999 carried out an experiment with a model turbofan in the Large
Low Speed Facility of the DNW
-
LLF (Figure 2.3). This test delivered a valuable database

10

on fan noise generation, in
-
duc
t propagation and far
-
field radiation (
kR
max

= 10.0). UTC
carried out a lined flow duct experiment on a locally and non
-
locally reacting liner. The
acoustic fields at either sides of the liners were determined using a traversing pressure and
velocity probe
. Measurements have been conducted for an axial mean flow velocity of 20
m/s and
kR

values ranging from 2.0 to 8.0 and will be compared with predictions of the
coupled FEM/BEM method. Rolls
-
Royce and ISVR carried out several no
-
flow
experiments on complex
duct geometries as turbine humps and buried exhaust cones. The
use of a flexible non
-
locally reacting liner proved successful and allowed the complex
geometry’s to be wrapped in a simple and cost effective manner for the lined builds. The
results are used
for a first assessment on the influence of exhaust geometry on fan noise
radiation.


2.4.4.
Concluding remarks


The project was successful in the development and validation of various duct
acoustics models and the execution of three validation experiments.

However at the end
of the project, it seemed that two industrial requirements could not be fully covered:



The inclusion of 3
-
D non
-
uniform flow and liner geometry’s.



The ability to calculate the 3
-
D in
-
duct propagation and radiation at realistic
frequenci
es (kR values ranging from 20 to 40, which are characteristic for 1 to 2
BPF of a modern turbofan).

Further developments and validations of the duct acoustics models towards more realistic
nacelle conditions are required before the models can be used as ma
ture industrial design
tools.

RIENSTRA
GALBRUN
MUNGUR
Figure 2.2 Comparisons results slowly varying duct: coupled FEM/BEM (Galbrun) and
analytical benchmarks (Rienstra and Mungur).


11



Figure 2.3 Test of NLR model turbofan in the DNW
-
LLF.


[By Edward Rademaker, NLR Department of

Aeroacoustics, The Netherlands]


3. EUROPEAN PROGRAMS STARTED IN 2001


3.1. SILENCE(R) (Significantly Lower Community Exposure To Aircraft Noise)


On 1 April 2001, the largest ever European aircraft noise research program, called
SILENCE(R), has been laun
ched. A consortium of 51 companies will collaborate during
4 years to validate noise reduction technologies that will allow as of 2008 quieter aircraft
operations by up to 6 decibels. The program is part of the 5
th

framework program of the
European Commiss
ion enabling a 50% funding of the total budget of more than 110
million Euro. The kick
-
off meeting took place in Copenhagen, from the 17
th

to 19
th

of
April and finished with a public session on 20
th

April involving presentations to the
European Commission
and major industrial partners on the objectives and technical
content of the project.


SILENCE(R) will address the issue of aircraft noise, a major cause of concern around
European airports, through three major objectives:



Large scale validation of nois
e reduction technologies whose development was
initiated by the European Commission and National projects in 1998.



Assessment of the applicability of these technologies to current and future
European products with minimum cost, weight or performance pena
lty.



Determination of the associated achievable noise reduction.


12


Novel concepts to be validated include low
-
noise fans, LP turbines, scarfed intakes,
novel intake, bypass and hot
-
stream liners, nozzle jet noise suppressors, active control
techniques and

airframe noise reduction technologies.


The SILENCE(R) program is linked to X
-
Noise, a European Thematic Network on
External Aircraft Noise. Participating companies consist of aircraft, aero
-
engine and
nacelle manufacturers, supported by research establis
hments and high tech engineering
companies and SMEs (see http.www.x
-
noise.net/ SILENCER kickoff PR pdf).


For information contact the X
-
Noise/SILENCE(R) communication manager
Dominique Collin at
dominique
.collin@x
-
noise.net

.


3.2. TURBO
-
NOISE CFD

3.2.2.
Introduction


This programme of research, which commenced in March 2000, is supported by the
European Union and is divided into four Workpackages. Workpackage 1 identifies the
key features of CFD codes th
at are required to model each type of turbomachinery noise
source, including blade row transmission effects. This covers steady and unsteady flows,
viscous and non
-
linear effects, boundary conditions, non
-
dispersive propagation and
meshing. The methodology

developed will be benchmarked against analytical results and
a feasibility study will also be conducted into the CFD modelling of fan broadband noise
sources.


In Workpackage 2 methods are being developed to link the source noise prediction
methods, devel
oped in Workpackage 1, with propagation models to carry the noise from
the source to the far field.


Workpackage 3 will benchmark the methodology developed above against
experimental data from:



Rotor/Stator viscous wake interaction



Potential interaction



Fa
n rotor

alone and buzz
-
saw tones



Rotor and Stator transmission and reflections.


13


Workpackage 4 will prove and refine the new methodology through a case study of
low noise design concepts identified in RESOUND and elsewhere, to develop the tools in
a work
ing environment prior to exploitation and to recommend low noise concept
improvements.


[By A. J. Kempton, Rolls
-
Royce plc, Derby, UK]


3.3. JEAN (Jet Exhaust Aerodynamics and Noise)


A research contract to develop new techniques for the prediction of Jet

Noise has been
awarded under the European Commission’s Competitive & Sustainable Growth or 5
th

Framework Programme. The JEAN project will be for three years and will be co
-
ordinated by Professor John Fitzpatrick and Dr. Craig Meskell of the Mechanical &
M
anufacturing Engineering Department, TCD. The work involves 13 partners from
across Europe representing the University, Research Centre and Industry sectors.


The objectives of JEAN are to develop methodologies for the prediction of noise
generated by jet
s including the effects of mixing enhancement and co
-
axial
configurations. The long term aim is to provide for design tools for the development of
low noise nozzles for HBR engines.
The specific technical objectives of the project are



to identify and de
velop optimal CFD methods for the calculation of the velocity
characteristics of jet flows of relevance to aircraft technology;



to develop aeroacoustic methods which use the CFD results as input for the
prediction of acoustic fields generated by exhaust
flows;



to validate the prediction techniques thus developed;



to identify the optimum prediction methodologies for particular applications.


The principal objectives of this project are:



development of predictive tools to assess future jet noise reduc
tion techniques;



to quantify effects of flow distortion and co
-
axial configurations.


[By J. Fitzpatrick, Trinity College, Dublin]



14

4. AIRFRAME NOISE


4.1. AIRFRAME NOISE FROM HIGH LIFT WING SLATS


SOURCE MODELLING


Airframe noise constitutes one impo
rtant component of aircraft noise in the approach
phase. As an outcome of numerous (world wide) experimental airframe noise studies
Handley Page type slats were identified as one of the dominating sources of airframe
noise. Therefore both scale
-
model and f
ull
-
scale slat noise data


as obtained within
recent dedicated experiments


were analysed, aiming at the development of an
engineering source noise prediction model. As a result upper slat trailing
-
edge noise was
found to represent the dominating source
mechanism. Both local edge
-
flow turbulence
and slat noise levels turned out to increase with decreasing wing angle
-
of
-
attack


while
level maxima shift towards lower frequencies


due to a corresponding increase of the
rear slat cove vortex size. According
ly noise level spectra scale with the vortex dimension
which in turn (in a first approximation) can be replaced by the slat chord (
c
s
). Squared
sound pressures were found to increase with flow speed (
v
) corr
e
sponding to a
v
4.5
power
law which is slightly l
ess compared to the
v
5

law known from trai
l
ing
-
edge noise theory.
This difference is assumed to originate from superimposed unsteady mean slot flow
fluctuations (monopole source).


A non
-
dimensional representation of noise spectra from slats of vastly dif
ferent scale
(Figure 4.1) exhibit excellent agreement, when based on a simple linear geometric scaling
law (SF = scale factor; R = radiation distance), inherently accounting for the linear
dependence of slat noise on wetted trai
l
ing
-
edge length.


Measured
slat noise polar directivities feature a shape similar to that of a compact
dipole with orientation perpendicular to the slat chord.


15
















Figure 4.1 Comparison of non
-
dimensional wing high lift far field slat noise spe
c
tra from
1/7.5 model
-
sca
le and full
-
scale wind tunnel experiments.


[By Werner Dobrzynski and Michael Pott
-
Pollenske,) DLR, Institut für Aerodynamik und
Strömungstechnik, Braunschweig/Göttingen, Germany]


4.2.
AIRFRAME NOISE REDUCTION BY ACTIVE FLOW CONTROL


Airframe noise contri
butes significantly to the total sound radiated during landing
approach. One dominant source is situated at the side
-
edge of the extended wing
-
flap.
Oscillation of the vortical structure at the flap side
-
edge leads to pressure fluctuations at
the rigid sur
face and thus to sound radiation into the far field [4.2.1], [4.2.2]. The
objective of the present study is to reduce this noise by blowing air into the vortical
structure. The basic idea behind this concept is to displace or destroy the vortical
structure
, reducing the surface pressure fluctuations and thus the amplitude of the radiated
sound.


In order to investigate the physical mechanism in detail, experiments on a swept
constant chord half
-
model were carried out in a wind tunnel with closed test sectio
n. The
angles of incidence and the flap
-

and slat angles of the model can be varied. At the flap
side
-
edge, air can be blown out with different velocities. Firstly, flow visualization with
Slat / Flap: 27°/39°

SF = 1/7.5
Scaled

Complete
Model


Full Scale

Wing Se
c
tion

SF/R = constant

Compl
ete scale model
ai
r
craft

Full scale wing se
c
tion


16

particle image velocimetry (PIV) was performed to investigate the f
low field at the flap
side
-
edge. The spatial vorticity distribution with and without blowing was measured.
Secondly, the change of the sound pressure level (SPL) due to the blowing was measured
with a single microphone. Furthermore, the distribution of the

sound sources on the
model was measured with a microphone
-
array.


The PIV measurements without blowing yield a rather complicated unsteady vortical
structure which confirms the assumption of a noise source at the flap side
-
edge. The
reason for the unstead
y motion of the vortices is the special geometry of the flap
-
wing
configuration. The swept main wing generates disturbances that influence the flow field
at the flap side
-
edge, thus generating vortex instabilities. This mechanism can also be
expected on a
wing of a real commercial aircraft. The measurements with the microphone
array show that the flap side
-
edge noise is present over a broad range of frequencies in
agreement with [4.2.3]. The angle of incidence determines if the slat
-

or the flap
-

noise is
d
ominant.


The vortical structure can be almost completely dispersed by the blowing and the
maximum vorticity in the vortex core is reduced. Furthermore, the SPL measured with a
single microphone is reduced by 3 to 4 dB above 1.25 kHz.


[By L. Koop, K. Ehre
nfried, A. Dillmann TU Berlin; and U. Michel, DLR Berlin]


5. DUCT ACOUSTICS

5.1. A CLASSIFICATION OF LINED FLOW DUCT MODES


For the relatively high frequencies relevant in a turbofan engine duct the modes of a
lined section may be classified in two categ
ories: genuine acoustic 3D duct modes
resulting from the finiteness of the duct geometry, and 2D surface waves that exist only
near the wall surface in a way essentially independent of the rest of the duct. The number
and location of the surface waves depe
nds on the wall impedance Z and mean flow Mach
number (see Figure 5.1).


The figure shows a typical case, corresponding to acoustic modes in a cylindrical
lined duct with uniform mean flow (
m
= 1, Helmholtz number = 5, Mach number = 0.5).
Displayed are tra
jectories in the complex plane of the eigenvalues (reduced axial wave
numbers) as a function of wall impedance
Z
. Re(
Z
) = 0.5 is kept fixed, while
Im
(
Z
) varies
from


to



. It is seen that most modes are genuine acoustic. They start and finish at a

17

hard
-
wall value (when |
Z
| =

), and stay near these hard
-
wall values for finite
Z
. Four
modes, however, move away from these hard
-
wall values, and become of surface type
(their existence physically is limited to the neighbourhood of the wall). Two of these
surf
ace waves (located outside of the “egg”) disappear to infinity when
Im
(
Z
)





.


Figure 5.1 Trajectories of reduced wave number (
m

= 1,


= 5) for
Im
(
Z
) varying from


to




and fixed Re(
Z
) = 0.5. The 2D surface wave limits are seen to be a very close
a
pproximation of the actual duct mode. The “egg” separates the regions of occurrence of
the surface waves.


[By S.W. Rienstra, Eindhoven University, The Netherlands]


5.2. ON THE AZIMUTHAL MODE PROPAGATION IN AXISYMMETRIC DUCT
FLOWS


One of the main sound s
ources in aeroengines is the tone noise of the rotor stator
interaction. CAA is part of a three
-
stage procedure that is developed to predict the sound
audible in the far field. The tone noise is generally dominated by only a few cut
-
on duct
modes, at the b
lade passing frequency and it's harmonics. It is much more efficient for
numerical computations to treat these few two
-
dimensional azimuthal Fourier
components (m). Therefore, under the assumption of axisymmetric mean flow without
swirl and with axisymmetr
ic acoustic boundary conditions, the three
-
dimensional
linearised Euler equations in cylindrical coordinates, can be decomposed into a Fourier
series in the azimuthal direction. This involves less restriction than the finite element
method (FEM) and the mu
lti
-
scale method (MS) assume for their solution. The
computational aeroacoustic (CAA) procedure applied for the time and spatial

18

discretisation is based on a seven
-
point 4th
-
order Dispersion
-
Relation
-
Preserving (DRP)
scheme and a 2N storage form Low
-
Dissip
ation and Low
-
Dispersion Runge
-
Kutta
(LDDRK) scheme respectively. Appropriate boundary conditions are prescribed for the
solid wall, inflow and outflow boundary [5.2.1].


The approach based on the Fourier decomposition is employed to the duct geometry
of R
ienstra and Eversman [5.2.2], similar to a generic aeroengine inlet. The results (
m

=
10,
n

= 1,


= 16) are in reasonable agreement with FEM and MS without as well as with
Euler based mean flow (
Ma

=


0.5). Although the CFD calculated flow field is diffe
rent
from the FEM potential flow and the MS assuming a 1D potential flow, the CAA
calculated sound field agrees rather well with FEM and MS. At the same time CAA is
much more applicable for future problems, and the CAA results also provide a good basis
for

the far field integration using the approach based on Ffowcs
-
Williams and Hawkins,
that also is in development.









Figure 5.2 (a) Comparison of normalised pressure contours by CAA, FEM and MS,
m

=
10,
n

= 1,


= 16, and
Ma

= 0.






Figure
5.2 (b) Comparison of normalised pressure contours by CAA, FEM and MS,
m

=
10,
n

= 1,


= 16, and
Ma

= 0.5.


[By C. Schemel, F. Thiele, TU Berlin; U. Michel, DLR Berlin; and X. Li, BUAA Beijing]



19

5.3.
BUZZ
-
SAW NOISE


One of the most significant sources o
f fan tonal noise from modern high
-
bypass
-
ratio
aero
-
engines, at high power operating conditions, is “buzz
-
saw” noise (also known as
“combination tones” or “multiple pure tones”). Aero
-
engines operating at supersonic fan
speeds generate an acoustic signatu
re in the forward arc containing energy at harmonics
of the engine shaft rotation frequency, known as Engine Order (EO) tones. These tones
are generated by the steady (in the frame of reference) “rotor
-
locked” pressure field
attached to a ducted fan. At su
bsonic fan speeds this pressure field is “cut
-
off” and decays
evanescently upstream of the fan, however at supersonic fan speeds the “rotor
-
locked”
pressure field propagates upstream of the fan, sweeping around the duct in a helical path.
The acoustic sign
ature measured close to the fan will be dominated by energy at the
Blade Passing Frequency (BPF) harmonics. However the energy in the EO harmonics
will be redistributed amongst these tones during the
nonlinear

propagation of the pressure
field. This leads
to a more ‘ragged’ and low
-
pitched noise referred to as the buzz
-
saw
signature of a supersonic fan.


Recent work, carried out at the ISVR on the X
-
Noise EC projects RESOUND and
DUCAT, has been concerned with prediction methods regarding buzz
-
saw noise, wit
h
particular emphasis on the prediction of buzz
-
saw noise in acoustically lined engine inlet
ducts. A simple, ‘engineering’, numerical method has been developed which calculates
the nonlinear propagation of the “rotor
-
locked” pressure field in a lined inle
t duct.


Also on the RESOUND project, Rolls
-
Royce plc carried out a model fan test (with a
lined inlet duct), and generated a large experimental database containing measurements
over a wide range of operating speeds. These measurements include many in
-
duc
t “buzz
-
saw” frequency spectra, and have provided an excellent source of experimental data to
benchmark, and further develop, the numerical prediction scheme.


The effect of an acoustic liner on the “buzz
-
saw” tones is largely dependent on the fan
operati
ng speed. At fan speeds slightly above sonic the liner is predicted to significantly
attenuate a broad band of the EO “buzz
-
saw” tones. However as the fan speed increases
the predicted attenuation falls, and there is less benefit obtained from the acoustic

liner.
Comparisons between measurement and prediction of “buzz
-
saw” noise show
encouraging agreement.


[By A. McAlpine, Institute of Sound and Vibration Research, University of Southampton,
UK]


20


6. THEORETICAL METHODS


6.1.
SOUND RADIATION OF THE VORTEX F
LOW PAST A GENERIC SIDE
-
MIRROR MODEL


The noise generated by the flow over blunt bodies and its propagation to the far field
plays a major role in many engineering applications. Presently various concepts are under
consideration to calculate the sound radi
ation into the far field such as combination of
flow field simulation and sound radiation integration. Based on the Navier
-
Stokes
equations the unsteady vortex shedding past blunt bodies can be numerically simulated.
Using the pressure fluctuations on the
body, the noise radiated to the far field is obtained
by an integral solution of the equation of Ffowcs Williams & Hawkings (FWH) [6.1.1].
The investigations performed are related to the unsteady flow around a side
-
mirror model
placed on a flat plate which

leads to the formation of trailing vortices and pressure
fluctuations on solid surfaces.


Unsteady Reynolds
-
Averaged Navier
-
Stokes calculations (URANS) allow the
computation of high Reynolds number turbulent flow in engineering applications such as
the si
de mirror but have difficulties to compute unsteady coherent structures due to a
misrepresentation of the interaction between turbulent/transient phenomena. The
predictive inaccuracy is particularly pronounced in conjunction with high frequency and
broad
-
b
and phenomena. Since the wall
-
resolving Large Eddy Simulation (LES) remains
rather unfeasible for high Reynolds number flows in the near future, Detached Eddy
Simulation (DES) [6.1.3] has recently become an attractive tool to investigate
aerodynamic flows.

DES avoids the high near
-
wall resolution by applying the RANS
approach in the vicinity of the wall and a modified Spalart
-
Allmaras one
-
equation
turbulence model in the far
-
field. In the overlap region, DES blends from a statistical to a
subgrid
-
scale mode
l without the use of shape functions.


For the validation of the FWH integration procedure the radiation of a sound source
into the far field is considered in a first step. For this purpose, an elliptical surface
encloses monopole and dipole sources, which

provide the boundary conditions for the
FWH procedure for field integration. Compared to the analytical solution the FWH
results are in excellent agreement with the pressure perturbation.


The capability of unsteady flow simulations in providing accurate
pressure
fluctuations on body surfaces is investigated in a second step for the complex unsteady

21

vortex flow over a generic side
-
mirror model [6.1.2]. Numerical simulations have been
performed for various grid arrangements, approximation schemes and modell
ing
approaches (URANS, DES). The results of the mean pressure distribution, the fluctuating
pressure levels as well as the surface pressure spectra are in fairly good agreement with
the experimental data available (Figures 6.1 and 6.2). Compared to URANS,
the results
for the sound spectra demonstrate that the DES
-
FWH approach is more suitable to
capture the main features of noise generation and sound propagation to the far field.


Figure 6.1 Measured and calculated surface pressure spectra at a selected se
nsor location
based on DES.


Figure 6.2 Measured and calculated sound spectra at a selected position.


[By D. Eschricht, J. Yan, F. Thiele, TU Berlin; and T. Rung, Bombardier, Hennigsdorf]


22


6.2. AEROACOUSTIC COMPUTATIONS OF UNSTEADY FLOWS USING CFD
AND AC
OUSTIC ANALOGY


The reduction of sound emission gains increasing importance in a variety of technical
fields such as fluid machinery construction. Expensive after
-
design action can be avoided by
applying pre
-
design sound prediction methods. In the field of

climatisation systems and fans
current prediction methods are mainly empiric or half
-
empiric solutions. Our present research
is focused on a numerical prediction method based on Computational Fluid Dynamics (CFD)
and Acoustic Analogy.


The method is imple
mented into the institute‘s research CFD solver. The parallelized,
compressible solver is able to do 3
-
dimensional, unsteady Reynolds Averaged Navier
-
Stokes
(RANS) as well as Large Eddy Simulations (LES). Algebraic, one
-
equation, two
-
equation
models can be

chosen for RANS calculations, the Smagorinsky, the Function
-
Structure or a
new two
-
equation model for LES. The acoustical module is based on the Ffowcs
-
Williams
Hawkings equation. It extracts the necessary information out of the unsteady CFD
computation,
performs the Ffowcs
-
Williams Hawkings integration and computes time
dependent density and pressure fluctuations as well as the related frequency spectra. The
acoustical module is steadily optimized concerning computing resources, e.g. CPU time and
memory n
eeds. The acoustical computations within each time step require presently an
overload of about 12% of memory compared to the pure CFD computation, whereas the
additional CPU time per time step is negligible.


Presently the acoustical module is tested with
a turbulent, unsteady flow around a 3
-
dimensional circular cylinder. The corresponding mesh consisted of about 3 million cells.
In the laminar case the sound emission results from the von Karman vortex street and is
dominated by pressure fluctuations induc
ing mainly lift and drag force fluctuations and
emitting a dipole sound field with its specific directionality. The third
-
octave spectrum
(Figure 6.3) was calculated from a turbulent flow of Re


59,000 using unsteady RANS.
The main peak was expected accor
ding to the observer position and the directionality at a
frequency of about
f
= 1170 Hz. Compared with experiments the computed main
frequency is slightly higher. The computed main peak is higher than the experimental
peak but agrees fairly well with the
one theoretically predicted for these experiments.



23


Figure 6.3 Third
-
octave spectrum of the sound emission of an unsteady flow around a circular
cylinder (Re


59,000).


[Dipl.
-
Phys. I. Pantle, Dr.
-
Ing. F. Magagnato, Prof. Dr.
-
Ing. M. Gabi, University o
f
Karlruhe, Germany]


6.3.
RESEARCH ON OPEN


AND DUCTED
-
ROTOR NOISE USING CAA
-
MULTIDOMAIN METHOD


The prediction of open or ducted rotor noise is a challenging problem as there are
large difference in scales between the noise source near the rotor and the

sound field. The
sound field generated by an open rotor and that radiated by the same rotor placed inside a
semi
-
infinite duct is simulated at DLR using a CAA multi
-
domain method. The linearized
Euler equations in cylindrical coordinates are used as gover
ning equations in solving this
problem. Both Tam's radiation and PML far field boundary conditions are used. A multi
-
domain Cartesian grid system is used so that the number of grid points can be kept as a
minimum. A combination of DRP and a specially optim
ized cell
-
centered high order
differencing scheme is implemented in the area of two Cartesian grid interfaces. Figure
6.4 gives the directivity obtained computationally in comparison with the exact solution
for an open rotor with rotational speed of
85
.
0


. There is good agreement between the
numerical simulation and the analytical solution. Figure 6.5) shows the pressure contour
radiated from the open end of the duct at
40

t

for the case
5
.
0

M

and its e
nlarged plot
close to the grid interface. It was demonstrated that the effect of a mean flow in the duct
may cause the propagation of acoustic modes which are cut
-
off for vanishing mean flow.
The results also prove the numerical treatment is successful. Th
e effect of mean flow
including a shear layer on the noise radiation off the ducted rotor is studied.


24



D
0
5
0
1
0
0
1
5
0
0
1
E
-
0
6
2
E
-
0
6
3
E
-
0
6
4
E
-
0
6
5
E
-
0
6
6
E
-
0
6
A
s
y
m
p
t
o
t
i
c
N
u
m
e
r
i
c
a
l
(
P
M
L
B
D
)
N
u
m
e
r
i
c
a
l
(
T
A
M
B
D
)

Figure 6.4 Comparison of the numerically obtained directivity at
R

= 15.0 and the exact
solution for


= 0.85 with two different far
-
field (PML and TAM
) boundary conditions
(BD).

x
r
-
2
-
1
0
1
2
0
0
.
2
0
.
4
0
.
6
0
.
8
1
1
.
2
1
.
4
d
u
c
t
w
a
l
l
(
b
)
x
r
0
5
1
0
1
5
0
2
4
6
d
u
c
t
w
a
l
l
M
M
=
0
.
5
P
M
L
P
M
L
P
M
L
d
u
c
t
w
a
l
l
M
M
=
0
.
5
P
M
L
P
M
L
P
M
L
d
u
c
t
w
a
l
l
M
M
=
0
.
5
P
M
L
P
M
L
P
M
L
(
a
)
t
=
4
0

Figure 6.5 Pressure contour at
t

40 and its enlarged plot close to grid interface; ducted
rotor case with
M

= 0.5 and


= 1.15.


[By J. Yin, J. Delfs, DLR, Braunschweig, Germany]


25


6.4.
AEOLIAN TONE SIMULATION USING HYBRID CFD
/CAA METHODS


A hybrid CFD/CAA method is validated for the problem of a cylinder in a cross
-
flow
at
M


= 0.3 and Re = 200. In the first step of the hybrid method the unsteady
compressible flow field in the immediate vicinity of a sound producing geometry i
s
computed. The acoustic field is computed with acoustic perturbation equations (APE) in a
second step using sources determined from the unsteady compressible flow field. The
APE system has been derived to solely describe the propagation of the acoustic mo
des.
Due to the excluded vortical eigenmodes the excitation of instabilities in global unstable
mean flows is prevented. Furthermore, the separation of the analysis of the flow field and
the acoustic field offers the possibility to take advantage of the di
sparity of the turbulent
and acoustic length scales at low Mach numbers. The left
-
hand side of the APE system
can be shown to be equivalent to the wave operator of Pierce [6.4.1], which agrees with
the linearized operator of Möhring's acoustic analogy.


A
high resolution CFD simulation of the unsteady compressible flow around a
circular cylinder was carried out as a reference solution and as a basis to evaluate source
terms for acoustic perturbation equations. The CFD simulation was based on an O
-
grid
with
a coarsest resolution of approx.~20 points per wavelength (PPW) at the outer
circumferential. The acoustic simulation was carried out on a coarsified O
-
grid with
approx. 8 PPW far
-
field resolution since the high order low dispersion and dissipation
CAA met
hods used for the discretization of the acoustic perturbation equations have
clearly improved limits for the highest resolved wavenumber. With the same radial
extension of r/d=80 cylinder diameters, the acoustic grid has 8 times less points.


Figure 6.6 de
picts a snapshot of the acoustic field, obtained from the CFD simulation.
Figure 6.7 shows the result using the APE system. It can be seen that no vortex street
occurs in the wake of the cylinder. Both simulations take into account convection effects,
whic
h can be seen qualitatively due to the two lobes tilted in upstream direction.



26


Figure 6.6 Unsteady perturbation pressure from high resolved CFD simulation.



Figure 6.7 Unsteady pressure field using acoustic perturbation equations (APE) with
source te
rm.


[By R. Ewert, M. Meinke, W. Schröder, Aerodynamisches Institut, RWTH Aachen,
Germany]


27


6.5.
ANALYTICAL PREDICTION OF BROADBAND NOISE RADIATED BY A
DUCTED FAN

The main difficulties for predicting fan broadband noise arise from the facts that there are
several competing noise generation mechanisms, and that input data are generally poorly
known. Present study has been performed within the framework of the European project
RESOUND
(Reduction of Engine Source Noise through Understanding and Novel
Design)
.


The method proposed here aims at avoiding these drawbacks, and at being fast and
easy to implement on a PC. It is devoted to random
-
dipole radiation from a ducted rotor.
It starts from an assumed flat blade loading spectrum, according to previous tests i
n a
SNECMA facility.

This hypothesis is also physically consistent with ECL tests made on a
low
-
speed cascade of highly
-
loaded airfoils to investigate its unsteady aerodynamic
behaviour.

In
-
duct sound pressure is computed using the Ffowcs Williams and Hawk
ings
equation in which the free
-
space Green's function has been replaced by the Green's
function in a cylindrical hard
-
walled duct. Free
-
field radiation is derived using the Tyler
and Sofrin model.


Applications are based on the data of the fan tested by R
olls
-
Royce in its anechoic
Ansty Noise Compressor Test Facility (ANCTF) during a previous European project
FANPAC. An example of the computed sound pressure levels in the fan forward arc, at
approximately 50% of the fan design speed, is shown in Figure 6.8
. The theoretical model
predicts good agreement between the in
-
duct and free field sound power levels if one
neglects the mean flow in the duct. These comparisons are shown in Figure 6.8 for a
range of fan speeds together with the measured forward arc soun
d power levels. The
sharp rise in the latter at transonic speeds (abscissa around
-
1) is due to the fact that the
multiple pure tones could only be partially separated from the broadband noise
component in the measured spectra. Subsequent FANPAC tests, in
which an acoustic
liner effectively absorbed these buzz
-
saw noise tones, showed that the broadband level
then varied continuously from subsonic to supersonic fan speed in accordance with the
present model.


Spinning mode content of broadband noise computed

at a given acoustic frequency
has also been analysed to better understand how sound waves propagate inside the duct
and radiate into the far field. Finally, the effect of the radial turbulence length scale along

28

the blade span can be predicted, and well d
uplicates an approximate closed form
expression derived by Mugridge for an airfoil in a turbulent flow (Figure 6.9).

0
30
60
90
0
2.5
5
7.5
10
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
5058 rpm
Acoustic Frequency
(kHz)
Radiation Angle (deg)
SPL (dB)


-4
-3
-2
-1
0
10 log
10
(RPM / Design RPM)
OAPWL (dB)
Down./Duct
Down./Free
Upstr./Duct
Upstr./Free
Test (Upstr.)
V
5
Law
5 dB

a) Upstream radiation at 5058

rpm

b) Overall sound power level versus rotation speed

Figure 6.8 Computation of upstream and downstre
am broadband noise radiation: In
-
duct
flow neglected, Drag/Thrust

=

0.398.

105
110
115
120
0.1
1
Length Scale / Span,

/
b
OAPWL (dB)
Upstream/ Duct
Upstream/ Free
Analytic (Mugridge):
W

b












b
e
b
W
b



1
1

Figure 6.9 Effect of radial turbulence length scale on upstream radiation at 5058

rpm: In
-
duct flow neglected, Drag/Thrust

=

0.398.


[By S. Lewy, ONERA]


29


6.6.
AEROACOUSTIC COMPUTA
TIONS OF INTERACTION NOISE RADIATED
BY FAN INLETS


Discrete
-
frequency tones due to wake interaction between rotor wake and outlet guide
vanes (OGV) are of particular concern in the design of advanced fans. For this reason,
CFD and CAA are strongly required

for modeling interaction noise problems. Based on
existing codes available at ONERA, a computation chain is at present investigated in the
framework of the European Project TurboNoiseCFD.


CFD RANS codes are expected to provide the pressure disturbance fi
eld upstream of
the fan, or the unsteady blade loads, required respectively by each of the two acoustic
models proposed. These are semi
-
analytical based routines, interfaced to a commercial
BEM scattering code. For both methods, the time
-
dependent acoustic

problem is
transformed into the frequency domain, and acoustic solution is split into spinning modes.


In the first model, the analytical solution of convected Helmholtz equation in an
infinite annular duct is used to match the unsteady fluctuating pressu
re computed by
CFD. The CFD solution is expanded into spinning and radial modes using a Hankel
-
Fourier transform, and converted into equivalent monopoles using an inverse method.
These adapted monopoles are then entered as input sources into the BEM code t
o
compute the radiated field solution of the scattering problem in the presence of the fan
inlet. The second model is based on the loading noise term of the Ffowcs Williams and
Hawkings equation, which gives the interaction noise due to the unsteady blade
loads, to
be provided by CFD. The analytical Green’s function of the infinite duct, commonly
used, is replaced here by the Green’s function of the fan duct, computed by BEM.


Although the present models are rather simple, they may be an alternative to more

advanced fan computations involving less restrictive assumptions for acoustics
(linearized
-
potential or linearized
-
Euler CAA solvers, direct Navier
-
Stokes/Kirchhoff
hybrid solvers), but still limited by high numerical difficulties. Validation of the
compu
tation chain is underway. It aims to be first applied to representative test cases,
modeling rotor/OGV configurations with simplified geometry, which have been tested in
the framework of DUCAT and RESOUND European projects.


[By C. Polacsek, ONERA, France
]




30

7. TECHNIQUES AND METHODS IN AEROACOUSTICS

7.1. LOCATION OF ROTATING SOURCES BY MICROPHONE ARRAYS


The applicability of phased arrays for acoustic source location has been extended to
the location of rotating sources, like broadband noise sources on h
elicopter and wind
turbine blades. A source reconstruction technique was developed that takes account of the
effects of source motion and Doppler frequency shift. Low signal/noise ratios can be dealt
with and the shear layer of an open jet wind tunnel does

not form a serious hindrance. The
technique was implemented in the computer program ROSI (“Rotating Source
Identifier”). [7.1.1]


ROSI was successfully applied to rotating whistles, blades of a helicopter in hover
and wind turbine blades. The test with th
e rotating whistles demonstrated convincingly
the capability to reconstruct the emitted sound. On the helicopter blades, rotating
broadband noise sources were made clearly visible. On the wind turbine blades, noise
emitted from the leading and trailing edg
e could be distinguished well.


An example of the difference between ROSI and conventional array software is
shown using array measurements on a model wind turbine in the DNW
-
LLF. Typical
results, at 8000 Hz, are shown in Figure 7.1 (conventional) and Figu
re 7.2 (ROSI). These
tests were set up to demonstrate (amongst others) the effects of trailing edge serrations on
the reduction of trailing edge noise. Using plots like Figure 7.2, it was shown that sound
radiation from the trailing edges was significantly

reduced. The leading edge source,
which can be seen near to the spinner on one of the blades in Figure 7.2, appeared to be
due to the junction of two different blade shapes.


Figure 7.1 Acoustic image of model wind turbine at 8000 Hz, obtained with conv
entional
array software.


31




Figure 7.2 Acoustic image of model wind turbine at 8000 Hz, obtained with ROSI.


[By P. Sijtsma, NLR, The Netherlands]



7.2. TURBINE BLADE/VANE INTERACTION NOISE: ACOUSTIC MODE
DECOMPOSITION IN UP TO 1500 RADIAL MODES


The sou
nd field in the outlet of a high
-
speed low
-
pressure turbine was studied. The
experiments aimed at a better understanding of the sound generating mechanisms inside
the three
-
stage turbine. Special attention was given to the acoustic impact of different
type
s of exit guide vanes (EGV) downstream of the turbine. Six radial rakes carrying ten
Kulite
-
sensor probes each were mounted downstream of the EGV’s in the cylindrical duct
section of the turbine exit. The rakes were traversed azimuthally over 180 degrees i
n steps
of 1.5 degrees to give a total of 240x30 measurement points. Acoustic measurements
were made at operating conditions from 63% to 99% rotor design speed resolving the
blade passing frequencies (BPF) of the three turbine rotors


The chosen experimen
tal setup allows the decomposition of the sound field into
azimuthal and radial modes for frequencies up to 6

kHz, Figure 7.3, with a total amount
of 1500 propagating radial modes. The results of the mode analysis permit the calculation
of the downstream r
adiated sound power. Using an extension of the theory of Tyler &
Sofrin [7.2.1], the dominant noise sources can be separated into rotor/stator
-

and
rotor/stator/EGV
-
interactions with associated azimuthal modes.


32


Figure 7.3 Instantaneous sound pressure pa
ttern in the measurement plane for BPF
1

at 6
kHz,
M

= 0.3.



The investigation was conducted in the framework of the Brite/Euram project
RESOUND (Reduction of Engine Source noise through Understanding and Novel
Design).


[By L. Enghardt*, U. Tapken*, W. Ne
ise*, F. Kennepohl**, K. Heinig** *DLR (German
Aerospace Center), Institute of Propulsion Technology, Department of Turbulence
Research, Berlin **MTU Aero Engines, Munich]


7.3.
STATISTICAL APPROACH TO TURBULENCE INDUCED NOISE AND
PHASED MICR
O
PHONE ARRAY M
EASUREMENTS


The noise prediction model SATIN (Statistical Approach to Turbulence Induced
Noise) was deve
l
oped at IAG (Institute of Aerodynamics and Gasdynamics, Un
i
versity of
Stuttgart) in the framework of SWING (Simulation of Wing Noise Generation), a Ge
rman
research project sponsored by DFG (Deutsche Forschungsgemeinschaft). SWING deals
with noise caused by high
-
lift devices of airplanes (airframe noise).


The formul
a
tion of SATIN is based on Lighthill’s Acoustic Analogy and allows
predicting both the fa
r field noise radiation as well as near
-
field wall
-
pressure
fluctuations. The latter is important for many practical problems because far
-
field noise

33

radi
a
tion may result from the scattering of wall
-
pressure fluctuations at geometrical
discontinuities.


T
he two building blocks of SATIN are a tailored Green’s function and a statistical
description of the turbulence. The first is computed for arbitrary geometry u
s
ing a
boundary
-
element method (BEM) and represents the influence of a solid body on the
sound fi
eld. The structure of the turbulence is defined via correlations of the turbulent
velocity fluctuations The required input values of SATIN are local properties of
turbulence, namely the turbulent kinetic energy and the integral length scale which can be
ob
tained from steady solutions of the Re
y
nolds
-
averaged Navier
-
Stokes equations or for
simple cases, e.g. boundary layers, from mea
s
urements.


Validation measurements were carried out using a phased microphone array (AAS). It
consists of 48 to 96 microphones
. Data acquisition, processing and visualization is
performed with a no
r
mal PC. The data can be recorded over all 96 channels continuously
over 15 s with a sampling rate of 48 kHz and 18 bit resolution.

Measurements of trai
l
ing
-
edge noise from a thin flat

plate were performed in the acoustic
wind tunnel at the Institute for Acoustics and Speech Communication of Dresden
University of Technology (AWD) (see Figure 7.4).


Figure 7.5 shows a good agreement between the predicted and measured sound
spectra. The
turbulence quantities were obtained from a RANS calculation.





Figure 7.4 Experimental set
-
up for
mea
s
urements of trailing
-
edge
noise of a thin flat plate (AWD).



Figure7.5 Measured and predicted sound spectra of a
thin flat plate (
U


=″㘮㌠洯猬s
c
f

= 0.0037,


=‴⸸浭⤮


[By
Jasmin S. D. Ostertag, Sandro Guidati, Gianfranco Guidati, Siegfried Wagner
Institute of Aerodynamics and Gasdynamics, University of Stuttgart,
Ge
r
many]


34


8. FLIGHT TESTING FOR NOISE

8.1. QUIET NOISE DEMONSTRATOR


Boeing and Rol
ls
-
Royce have completed a noise reduction flight test program, known
as the Quiet Technology Demonstrator (QTD), in which a Rolls
-
Royce Trent 800 engine
was modified with a package of noise reduction technologies developed collaboratively
by the two aerosp
ace companies. Using a Boeing 777
-
200ER

the three week flight
-
test
demonstrated noise levels significantly below those of a standard 777: takeoff jet exhaust
noise was reduced by up to four decibels and inlet fan noise was reduced by up to 13
decibels. Eng
ineers used saw
-
tooth
-
shaped aerodynamic devices at the rear of the nacelle
and on the exhaust nozzle to control the mixing of the hot jet exhaust, the bypass stream
and the ambient air, the shape of the devices being determined by computational fluid
dyna
mics modelling and wind tunnel testing using scale models. Fan noise also was
reduced with extensive acoustic improvements to the redesigned engine nacelle inlet, in
which a new technology called Amax (area maximisation) increased by 30 percent the
area of

acoustic treatment in the inlet casing, and a new lining design was used that
reduces objectionable "buzz saw" noise passengers often hear during takeoff and climb.



Figure 8.1 The quiet technology demonstrator.


[By

A. J. Kempton, Rolls
-
Royce plc, Der
by, UK)



35

9. HELICOPTER NOISE


9.1. BREAKTHROUGH IN BVI NOISE REDUCTION OF HELICOPTERS BY
ROTOR ACTIVE CONTROL


The active blade root control system installed on a BO 105 helicopter has been
successfully tested in open loop configuration by EUROCOPTER DEUT
SCHLAND
GmbH (ECD). Flight tests have demonstrated the noise and vibration reduction potential
of the individual blade root control (IBC) technology. The succeeding investigations
therefore concentrated on the realization and the testing of a closed loop n
oise and
vibration control using IBC technology.


The BO 105 IBC demonstrator (see Figure 9.1) uses proven electro
-
hydraulic blade
pitch actuators with adequate authority for noise reduction from ZFL. This actuation
system is controlled by an embedded dig
ital computer in combination with high
performance signal processing equipment for the data transfer between the rotating and
non
-
rotating system from the DLR. For the exterior noise control a complex sensor
system is installed consisting of blade pressure

transducers and a landing gear mounted
microphone array.


The applied BVI noise control is based on a newly developed concept for minimizing
an appropriate BVI index using either blade pressure or microphone signals by applying a
2/rev IBC feedback. For t
he recent flight tests (performed Nov 01) only the microphones
installed on the landing gear were used for BVI noise detection and generation of the BVI
index
based on sound pressure signals
.
Some details of this closed loop control system are
presented in

Figure 9.2. A “Golden Section” algorithm was applied for the optimisation
of the 2/rev IBC phase angle towards the minimum of the BVI index. The IBC amplitude
was not optimised during the tests and was therefore kept at a constant value of 1°. This
approa
ch was in accordance with open loop flight tests which generally have indicated
that the IBC amplitude of 1° is most favourable for the reduction of the BVI noise
emission.
The evaluation of the flight tests in closed loop configuration is not finished at
present, but first results are very promising. For example major BVI noise reductions in
the range of 5
-
6 dB(A) could be achieved for the approach flight procedures with 6° and
8° slope angle (see Figure 9.3).



36



Figure 9.1 The BO 105 IBC demonstrator us
ed for flight tests.


Landing Skid
Micros
Rotor
IBC Actuators
Micro Signal
IBC PHASE
ANGLE
Signal Processing -
BVI Index
BVI Index
BVI Noise
Controller
(Golden Section)


Figure 9.2
BVI noise control concept.



37

A
p
p
r
o
a
c
h
:
A
n
g
l
e
-
D
e
g
S
E
L
(
S
o
u
n
d
E
x
p
o
s
u
r
e
L
e
v
e
l
)
-
d
B
5
6
7
8
9
8
6
8
8
9
0
9
2
9
4
9
6
9
8
1
0
0
I
B
C
O
f
f
I
B
C
O
n
-
5
.
5
d
B
-
4
.
1
d
B
-
6
.
8
d
B
-
4
.
2
d
B


Figure 9.3 Reduction of sound exposure levels achieved by rotor active control for 6° and
8° approach flight procedures.


[By Dr Marius Bebesel / Dieter Roth EUROCOPTER DEUTSCHLAND GmbH, Mu
nich,
Germany]


10. COMMUNITY NOISE


10.1. DLR


PROJECT, QUIET AIR TRAFFIC


A CONTRIBUTION TO NOISE
REDUCTION THROUGH INTEGRATED APPROACH IN RESEARCH


Aircraft noise is an urgent environmental problem. It will become more and more
critical in the near fu
ture because the ongoing growth of air traffic will force an increase
of the noise load around airports which cannot be overcompensated by the noise
reduction techniques for airplanes currently available.


In the long term a technical potential of engine
and airframe noise reduction of about
10 dB is given by experts. However short and mid
-
term measures are necessary to reduce
noise as fast as possible. Such measures are the target of the DLR project “Quiet Air
Traffic”. The objective is to realize a noise

emission reduction of about 3 dB with respect
to the current state in the extended environment of civil airports. Five workpackages
which are linked with each other build an interdisciplinary research system covering

38

human
-
specific, technical/operational
as well as regulatory aspects of noise reduction.
The targets of the different workpackages are:



Definition of scientifically sound critierion for the assessment of nightly aircraft
noise as a basis for noise reduction measures and legislation (sleep lab
oratory and
field studies).



Development of construction instructions for the aircraft industry to reduce engine
and airframe noise (wind tunnel and engine test facility experiments, theoretical
investigations).



Improvement of currently used noise
-
abate
ment flight procedures for airlines, air
-
traffic control and airport authorities (simulation, flight experiments).



Development of concepts for regulatory noise reduction measures for legislative
and airport authorities.



Definition of improved aircraft
noise prediction procedures for environmental
planning and for modeling of noise reduction strategies (computer simulation,
validation experiments).


An effective abatement of aircraft noise cannot consist just in a technical and
operational reduction of t
he physically measurable sound. Much more it must take into
account the physiological and psychological effects of noise on people. In particular, our
knowledge of the effects of nightly aircraft noise is still rather poor and needs to be
considerably i
m
pr
oved.


The straightforward way to minimise aircraft noise on the ground is to reduce the
noise at the source. Since fan noise becomes more and more a dominating source with
respect to decreasing jet noise the activities focus on the design of an optimised
fan based
on the reduction of fan rotational speed. Additionally the use of active noise control for
the reduction of very disturbing discrete tones is investigated.


While engine noise has been reduced effectively during the last decades, airframe
noise h
as increased due to the use of modern high
-
lift devices. Major sources of airframe
noise are the landing gears, flaps and slats and the gear
-
wake/flap interaction. First wind
tunnel studies were performed to identify source mechanisms, prediction schemes a
nd
reduction technologies. Flyover tests were realised to validate the windtunnel results as
well as to generate airframe noise databases.


A further goal is to define operational quiet procedures with optimised altitude and
speed
-
profiles for takeoff and
landing of modern civil jet aircraft which can provide a

39

noise reduction around civil airports. In a next step these procedures have to be integrated
into modern air traffic management concepts.


As operational procedures transport political concepts for a
ircraft noise reduction are
used to reduce interest conflicts between people living around airports, carriers, airport
authorities and passengers. Possible measures are e.g. landing fees, noise quotas, steering
of modal split, more effective certification
regulations or airport co
-
operations.


Noise calculation procedures are an important tool for land
-
use planning and noise
legislation as well as for the validation of the effects of different noise reduction
strategies. An advanced simulation procedure bas
ed on a component source model (jet,
fan and airframe noise) is currently under development. Such a procedure allows one to
estimate a noise level time history at a given observer. Based upon that cumulative noise
load expressed in long terms “noise descri
ptors” can be estimated. Additionally an
advanced sound propagation model will be integrated. The approach is to couple
meteorological models with sound propagation models thus giving a consistent
description of the complex topography, atmosphere and sound

field.


All these activities are performed in close contact to carriers, industries, legislators,
agencies and associations as well as to universities and other research centers.



[By
Burkhard Gölling, Ullrich Isermann DLR, Institute of Aerodynamics and
Flow
Technology, Göttingen]


REFERENCES

4.2.1

J. C. Hardin 1980
AIAA Journal
. Noise radiation from the side edges of flaps
18
(5), 549
-
552.


4.2.2

C. L. Streett 1998
AIAA

Paper 98
-
2226. Numerical simulation a flap edge flow
field.


4.2.3

H. M. M. van der

Wal and Pieter Sijtsma 2001
AIAA

Paper 2001
-
2170. Flap noise
measurements in a closed wind tunnel with a phased array.


5.1.1

S. W. Rienstra 2001
7th AIAA/CEAS Aeroacoustics Conference, Maastricht, The
Netherlands
. A Classification of Duct Modes Based on

Surface Waves 2001
-
2180.


5.2.1

Li, Xd. Schönwald, N. Thiele, F.
AIAA/CEAS
2001. Numerical Computation of
Sound Propagation and radiation in a Duct 2001
-
2179.


40

5.2.2

S. Rienstra and W. Eversman 2001
Journal of Fluid Mechanics.
A Numerical
Comparison Betwe
en Multiple
-
Scales and FEM Solution for Sound Propagation
in Lined Flow Ducts
437
, 367
-
384.

6.1.1

F. Farassat 1975
NASA Technical Report R
-
451
.

Theory of noise generation from
moving bodies with an application to helicopter rotors.

6.1.2

R. Höld, R. Bren
neis and A. Eberle
AIAA
-
99
-
1896.

Numerical Simulation Of
Aeroacoustic Sound Generated By Generic Bodies Placed On A Plate: Part 1



Prediction of Aeroacoustic Sources.

6.1.3

A. Travin, M. Shur, M. Strelets and P. Spalart 2000
Journal of Flow, Turbulence
and Combustion.
Detached eddy simulations past a circular cylinder
63
, 293
-
313.

6.4.1

A. Pierce 1990
Journal of the Acoustical Society of America
. Wave equation for
sound in fluids with unsteady inhomogeneous flow
87
(6), 2292
-
2299.

7.1.1

P. Sijtsma, S. O
erlemans and H. Holthusen 2001
AIAA

Paper 2001
-
2167Location
of rotating sources by phased array measurements.

7.2.1

L. Enghardt, U. Tapken, W. Neise, F. Kennepohl, and K. Heinig 2001
7
th

AIAA/CEAS Aeroacoustic Conference
, Maastricht.
Turbine Blade/Vane
Int
eraction Noise: Acoustic Mode Analysis using in
-
duct Sensor Rakes,

CEAS/AIAA Paper 2001
-
2153.


ACKNOWLEDGEMENT

The authors are pleased to acknowledge the contribution of Mrs Susan Brindle to the
production of this paper.